Research On Hybrid Mean Center Particle Swarm Optimization And Its Application In Reservoir Optimal Operation

Developing and utilizing clean energy rationally is a major strategic decision for the country to achieve sustainable development.As a clean energy source,hydropower has been widely exploited.However,with the establishment of large-scale cascade hydropower stations,the joint operation of hydropower stations is increasing difficult.This paper focuses on the improvement of particle swarm optimization and explores the higher performance of particle swarm optimization.Taking the optimal operation of Shuibuya,Gaobazhou and Geheyan reservoir groups as the object,the improved particle swarm optimization was used as the optimizing tool to solve the established optimal operation model of reservoir.The main work of this paper is as follows:(1)The optimal operation model of reservoir based on particle swarm optimization is established.According to the basic theory of particle swarm optimization,the water level at the end of month of each reservoir is used as the decision variables,and the objective function with the largest annual total power generation is established.In order to speed up the solving speed,the penalty function mechanism is introduced to deal with the solution that does not satisfy the condition.(2)A hybrid mean center opposition-based learning particle swarm optimization is proposed.The new algorithm greedily selects the mean centers of all the particles and some high-quality particles,and obtains a high-quality hybrid mean center,which can perform a detailed search on the region where the particles are located.At the same time,the opposition-based learning of the hybrid mean center enables the particles to explore more new areas;the square wave control mechanism is introduced to further balance the exploration and exploitation capabilities of the algorithm.(3)A particle swarm optimization for dynamic spatial variability of superior quality individuals is proposed.Although the hybrid mean center is a high-quality particle constructed by combining population advantage information,the effect of single high-quality particles on increasing population diversity is limited.Therefore,the new algorithm selects some high-quality particles from the population,chooses two different dimensions from the high-quality particles,makes one dimension fly to the other dimension,obtains new variation dimension values,and the mutated dimension is reduced as the number of iterations increases,thereby enhancing population diversity.(4)A stochastic single-dimensional mutated particle swarm optimization based on dynamic subspace is proposed.The previous two algorithms all adopt the strategy of updating the overall dimension,because the one dimension or a few dimension often do not reach the optimal solution,resulting in the particle fitness value is worse.The new algorithm randomly and dynamically selects several dimensions of the particle to form asubspace,and randomly chooses one-dimension different from the subspace to mutate.In the early stage of evolution,most of the dimensions are used to form the subspace,which increases the diversity of the mutated dimension.Later,a few dimensions are selected to form the subspace,which enhances the ability of the particle to search fine.Finally,the improved particle swarm optimization is used to solve the optimal operation model of reservoir with the largest total power generation as the objective function.The experiment has achieved good results.